dsh-agnes-omni
Run the following command in DeepSeek Harness:
dsh plugin install wumu1111111/dsh-agnes-omni
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install wumu1111111/dsh-agnes-omni in your terminal to install the plugin; source code is available at https://github.com/wumu1111111/dsh-agnes-omni
About this plugin
Pure-text adapters like DeepSeek reject image content blocks outright, so a DSH agent cannot process a screenshot, a UI mockup, or a chart that a user pastes or drags into the conversation, let alone generate an image in return. dsh-agnes-omni was built to close exactly that gap: it wires the free Agnes omni-model into DSH so the agent gains both visual understanding and image generation, and a user simply drops a picture into the chat to trigger analysis.
Three components work together under the hood. agnes_vision calls agnes-2.5-flash for image comprehension, covering screenshot OCR, object recognition, and chart or UI analysis. agnes_image uses agnes-image-2.1-flash for text-to-image and image-to-image editing, saving results straight into the session workspace. A visual bridge intercepts any base64 or attachment image before it reaches the text adapter, decodes it to a local file, swaps in a path reference, and steers the model to call agnes_vision - the whole pipeline is transparent to the user: paste, send, get an answer.
It is a natural fit for DSH users who regularly need an agent to review a screenshot or design file, for solo creators who want quick image generation or light style transfer, and for anyone who wants to add multimodal capability to their workflow at zero cost with no hardcoded keys.
Use Cases
- Paste a screenshot into chat and let the agent auto-extract OCR text or analyze UI layout
- Describe an image in a single sentence and get it generated and saved to the session workspace
- Hand over a reference image path plus edit notes for a quick image-to-image style transfer
Best For
- Developers who regularly need a DSH agent to review screenshots, mockups, or charts
- Solo creators who want zero-cost multimodal capability without hardcoded API keys
- Technical users already building task agents on DSH and looking to fill the vision and image-generation gap
Related Plugins
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.
An armor-breaking plugin for DeepSeek: rewrites instruction priority, expands refusal-blocked phrases, persists across sessions, and shows a green active indicator.